Abstract
Songbirds use multiple stopover locations to rest and refuel for subsequent flights while migrating between breeding and non-breeding areas. Selection of high quality stopover habitat may allow migrants to minimize time and energy spent in migration and maximize fitness. Migrants should therefore attempt to select stopover habitat that affords them suitable safety, shelter, and food. A better understanding of habitat attributes that support high numbers of migrant landbirds during stopover is needed to develop conservation strategies for these species, many of which are in decline. I examined how migrant density during stopover in the Western Lake Erie Basin (WLEB) of Ohio was influenced by local- and broad-scale habitat variables, specifically: patch vegetation composition and structure, patch size and isolation, patch distance from the Lake Erie shoreline, patch distance from a river or stream, and wetland cover surrounding a patch. I used a generalized random tessellation stratified (GRTS) approach to select forested study sites within 22 km of the lakeshore along a 70 km stretch of shoreline between Toledo and Sandusky, Ohio, USA. Observers conducted over 800 point counts annually from mid-April through late May in 2011 and 2012 at a total of 135 locations. Point count data on Blackpoll Warbler (Setophaga striata), Black-throated Green Warbler (S. virens), and a guild of transient wood warblers (Parulidae) were analyzed using the Dail and Madsen (2011) generalized hierarchical N-mixture model. This is an open-population model that simultaneously estimates parameters that influence the abundance and detection probability of study species. Detection probabilities for transient migrants varied by survey technician, wind speed, and survey time, highlighting the importance of accounting for detectability in bird migration studies. Broad-scale variables such as distance to the lakeshore, patch isolation, and wetland cover were generally better predictors of migrant abundance than proximity to a river or stream or patch area, and local-scale variables (habitat structure and vegetation). Densities for both study species were greatest in forest patches near the lakeshore. For the transient warbler guild, densities declined about 3.4% per km from the lakeshore. Density of the transient warbler guild was greater at sites that had more emergent aquatic midges (Chironomidae) in 2012, suggesting that midges could help explain the distribution and abundance of migrants in the WLEB. While densities of transient migrants were greater near the coast, inland forests often supported more transient migrants per patch than coastal forest patches. Conservation efforts in the region should seek to protect, create, restore (1) forested areas adjacent the lakeshore and (2) any forest within 0.5–10 km of the coast, with priority given to forests that are larger (>20 ha) and closer to the lakeshore and wetlands. Results from this study should be important in developing conservation plans for migratory songbirds and for guiding decisions regarding local habitat management and the placement of wind turbines within the landscape.